arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.02483cs.CRcs.AIcs.CL
收藏

幽灵工具调用:投机性智能体工具的发布时隐私保护

Ghost Tool Calls: Issue-Time Privacy for Speculative Agent Tools

Max Planck Institute for Software Systems(马克斯·普朗克软件系统研究所);EPFL(苏黎世联邦理工学院);Aarhus University(阿arhus大学)

Bardia Mohammadi, Lars Klein, Akhil Arora, Laurent Bindschaedler

AI总结:

针对工具增强型语言智能体投机性预发调用泄露用户意图的问题,提出投机性工具隐私契约,在发布时而非提交后保护隐私。

摘要:

工具增强型语言智能体投机性地发出可能的未来工具调用以隐藏延迟,但这些调用在智能体提交分支之前将推断出的用户意图泄露给外部服务。每个收到调用的外部观察者在智能体放弃分支后仍保留该披露。问题在于时机,而非授权:提交后的清理、只读限制或访问控制白名单都无法撤回观察者已持有的信息。我们将这些调用称为幽灵工具调用,并提出投机性工具隐私契约,这是一种运行时抽象,将提交前的观察视为与状态突变不同的第一类效应。我们在原型运行时中实现了该契约,并在三个语料库上评估了十二种策略。投机性调度增加了观察者能够推断用户意图的程度;事后过滤器、只读限制和访问控制白名单无法消除这种推断;只有那些在调度前改变或抑制投机性调用的参数或目标投影的发布时策略才能减少这种推断。

英文摘要:

Tool-augmented language agents speculatively issue likely future tool calls to hide latency, but those calls leak inferred user intent to external services before the agent commits to the branch. Every external observer that received the call retains the disclosure after the agent abandons the branch. Timing is the issue, not authorization: no commit-time cleanup, read-only restriction, or access-control allow-list unsends what an observer already holds. We call these invocations ghost tool calls and propose Speculative Tool Privacy Contracts, a runtime abstraction that treats observation before commitment as a first-class effect, distinct from state mutation. We implement the contracts in a prototype runtime and evaluate twelve policies across three corpora. Speculative dispatch increases what an observer can infer about user intent; post-hoc filters, read-only restrictions, and access-control allow-lists leave that inference intact; only issue-time policies that change or suppress the speculative call's argument or destination projection before dispatch reduce it.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑